Minimal Clinically Important Difference and Substantial Clinical Benefit Using PROMIS CAT in Cervical Spine Surgery

Abstract

Study design: This was a prospective cohort study.

Objective: The objective of this study was to establish minimal clinically important difference (MCID) and substantial clinical benefit (SCB) thresholds for Patient-Reported Outcomes Measurement Information System (PROMIS) in cervical spine pathology.

Summary of background data: PROMIS enables improved psychometric properties with reduced questionnaire burden through computer adaptive testing. Despite studies showing good correlation with "legacy" outcome measures, literature on the clinical significance of changes in PROMIS scores is scarce.

Materials and methods: Adult patients undergoing cervical spine surgery at a single institution between 2016 and 2018 were prospectively enrolled. Patients completed questionnaires [Short Form-36 (SF-36), Neck Disability Index (NDI), Visual Analogue Scale Arm/Neck, and PROMIS Pain Interference (PI) and Physical Function (PF)] preoperatively and at 6 months postoperatively. MCID was calculated using distribution-based and SCB using anchor-based methods. The SF-36 Health Transition Item was utilized as an anchor with the cutoff values chosen using receiver operating characteristic curve analysis.

Results: There were 139 patients meeting inclusion criteria, with a mean age of 56.4 years and diagnoses of myelopathy (n=36), radiculopathy (n=48) and myeloradiculopathy (n=49). There were significant improvements in PROMIS PF, PROMIS PI, NDI, and SF-36 preoperatively to postoperatively (P<0.001). The test-retest reliability of all tests was excellent (intraclass correlation coefficients=0.87-0.94). PROMIS, SF-36, and NDI were all correlated with the anchor question (|r|=0.34-0.48, P<0.001). MCIDs were 8.5 (NDI), 11.1 (SF-36 Physical Component Score), 9.7 (SF-36 Mental Component Score), 4.9 (PROMIS PI), and 4.5 (PROMIS PF). SCB was 13.0 (NDI), 24.0 (SF-36 Physical Component Score), 11.8 (SF-36 Mental Component Score), 6.9 (PROMIS PI), and 6.8 (PROMIS PF). MCIDs were greater than standard error of measurement for all measures.

Conclusions: We report MCID of 4.9 (PI) and 4.5 (PF) and SCB of 6.9 (PI) and 6.8 (PF). These data support the use of PROMIS computer adaptive tests in cervical spine patients and provide important reference as PROMIS reporting becomes more widespread in the literature.